We showed previously that CYP1B1-null mice developed 10 times less lymphomas than wild-type mice after receiving 7,12-dimethylbenz[a]anthracene (DMBA). In this study a 10-fold lower dose was applied to differentiate between toxicity induced lymphomas (200 micro g/mouse/day) and tumor initiation (20 micro g/day). DMBA adducts to DNA of organs of mice, or to DNA of V79 cells expressing single mice or human cytochrome P450 isoenzymes were also measured. Mice were dosed three cycles of 5 days/week with DMBA in corn oil orally. Histopathology was determined at intermittent death or 1 year after dosing. DMBA-DNA adducts were assayed by (32)P-postlabeling. At 20 micro g/day, wild-type mice developed ovary (71%, stromal cells derived), skin (36%), uterus (64%) and lung (14%) hyperplasias. At this dose the CYP1B1-null mice developed no lymphomas, 25% ovary (epithelial cells derived), 8% skin, 58% uterus and 33% lung tumors. Oil control mice (n = 35) developed only eight, mostly different, hyperplasias. Wild-type mice had more DMBA-DNA adducts than the CYP1B1-null mice. The differences were highest in thymus, spleen, ovaries and testes (5-7-fold). Additionally, one specific DMBA-DNA adduct was reduced in CYP1B1-null mice. V79-cells expressed mouse CYP1B1 was 35 times more active than mouse CYP1A1 in forming DMBA-DNA adducts. Human CYP1B1 was 2.5 times less active than mouse CYP1B1 but 2.3-fold more active than human CYP1A1. CYP1B1 is the dominant enzyme in metabolizing DMBA to carcinogenic metabolites at high and low doses in mice, leading to an increased tumor rate of especially the ovaries at low doses of DMBA. Wild-type mice had more DMBA-DNA adducts than CYP1B1-null mice. Additionally, a specific adduct was less present in the CYP1B1-null mice. Human CYP1B1 was less active than mouse CYP1B1, but more active than human CYP1A1 in forming DMBA-DNA adducts. Thus, we expect CYP1B1 to be an important DMBA activating enzyme in humans also.